W3C

- DRAFT -

Accessible Platform Architectures Working Group Teleconference

14 Aug 2019

Attendees

Present
jasonjgw, Joshue, Joshue108
Regrets
Chair
jasonjgw
Scribe
janina

Contents


<Joshue108> trackbot, start meeting

<trackbot> Meeting: Accessible Platform Architectures Working Group Teleconference

<trackbot> Date: 14 August 2019

XR Accessibility: insights from the research literature.

<scribe> scribe: janina

jo: Looking at adding research references from other RQTF docs into XAUR

jgw: Continuing discussion of XR a11y, our lit review, and use cases + reqs that emerge
... We have AR and VR a11y docs, and these raise issues not currently addressed in WCAG
... e.g. how to provide descriptions of complex scenes that include many objects, which may in turn be made up of many objects
... Also issues of how to combine info of the real scene with the additional augmentation reality layer
... Next, the notion of sensory substitution which requires learning and aclimitization
... Recalling a demo that scanned a visual environment and used musical sounds to try and paint the scene aurally

jo: Agrees on the many aspects
... Concerned to pull in relevant items into XAUR

<Joshue108> https://docs.google.com/presentation/d/10ynaGMBAdiCLVyoyBDSNsNhtpQT9qm_QWO6VBI2LCGA/edit#slide=id.p

<jasonjgw> Janina notes the significantly enhanced execution time as the principal advantage of Web Assembly.

<Joshue108> JOC: DIscusses new approaches to memory management and API access in WebAssembly

<jasonjgw> The concern is that it may make debugging challenging.

<jasonjgw> Janina notes the potential advantages of being able to make efficient API calls from Web Assembly.

<Joshue108> JW: From what I've heard the API calls will be more efficient.

jgw: Believe the intent is to make api calls far more performant

<Joshue108> If they can make that easy we get better accessibiity support for Games and XR.

jgw: A11y support should also improve from within games, XR etc
... Presumes there's something a11y useful to call!
... Reminds me we also need to cover input methods

<Joshue108> Maybe WebAssembly would be the way to do that.

<Joshue108> JS: Am not convinced..

<Joshue108> JS: Think of people talking in a room, they can track this stuff well.

<Joshue108> People speaking etc can be identified etc

<Joshue108> Why not in XR?

<jasonjgw> Janina notes recent advances in room audio capture, including speaker identification. The same techniques could be applied in the VR enviornment. Example: directing one's voice at an object in the scene.

<Joshue108> There are things that have advanced - if you can direct a voice, why not gestures etc?

<Joshue108> JW: I think thats on the way.

<Joshue108> JS: Controllers are on the way out!

<Joshue108> lol

<Joshue108> JW; Well there is hardware appearing..

<jasonjgw> Janina notes that the performance issues addressed by WebAssembly are likely to be resolved.

<Joshue108> JS: There are temporary problems that are being fixed by things like WebAssembly..

<Joshue108> 10 years from now you wont need that..

<Joshue108> The optimisations are temporary.

<Joshue108> JW: They wont be using interpreted JS for sure, they will move to other languages etc

<Joshue108> But regarding WebAssembly we will learn from algos and machine learning stuff..

<Joshue108> We can offload some of this stuff on the hardware without remote service..

<Joshue108> Great performance leap, a la computer vision and other access needs.

<Joshue108> One objective should be getting a better understanding of where web based apps are going and how it relates to machine learning techniques etc

<Joshue108> Then we can get the requirements.

<jasonjgw> Janina: we're approaching the imit of human vision with video, especially with 8k resolution - diminishing returns beyond this point.

<jasonjgw> At some point we surpass the limits of human vision. The same may well be true of audio (at 192khz sample rates?)

<Joshue108> JW: Right my latest devices goes very high..

<jasonjgw> Janina suggests we should experience some of the high-quality audio available in suitably equipped cinemas.

jo: Even if asm is a stop gap, it's still a necessary step
... We need to remember to understand which is the platform we need to support, asm, gl, gpu??
... I'm currently liking asm

<Zakim> Joshue, you wanted to mention learnings and questions are glTF

jo: Recalls matthew mentioned as possibly being machine readable and support semantics

<Joshue108> JOC: I want to bring up also machine readable advantages of glTF also.

jgw: I'm unclear ... Need to validate assumption ... adding api to web platform ... are they then auto available to current mechanisms? With no extra work?
... to be available to interpreters or compiled code?

jo: That's the question. It needs to be done well
... These are performance related issues
... We're looking how to mainstream semantics

jgw: This also emphasizes need for AOM to proceed
... Good references to explain the transmission format would be helpful

jo: Notes that we're assuming AOM is our solution
... It has certain advantages for handshake and transmission of certain relationships, but we need a broader conversation with people working on these technologies

jgw: We need to tease out any additional reqs from XR

<Joshue108> JOC: I want to bring up also potential machine readable advantages of glTF also, waiting for more info from Matthew.

jgw: We need AT that we currently operate, so interface into XR using api probably
... Second is to take AT functionality and build it into XR apps directly
... There seem to be people interested in both approaches

<jasonjgw> Janina: suggests a good definitional discussion of XR.

<Joshue108> JW: I'm suggesting a related one that builds on that.

<Joshue108> I suggested this last week.. what are the features in XR that make it different from traditional applications?

jgw: Also, a conversation on what makes XR different from apps we're currently accustomed to, therefore what's missing in scenarios we currently understand for a11y
... Brings movement and movement tracking to mind that we need to explore
... Also find WebXR device api useful, assumptions clearly defined
... We wouldn't consider that exhaustive of what XR will provide, it's only a 1.0
... But, it's important

<Joshue108> JS: I mentioned this to Josh on Monday..

<Joshue108> Suggesting XR needs to move forward - is XAUR a 2.0 of the Maur or a seperate document.

<Joshue108> Depends on how different it is.

<jasonjgw> Janina: raises the question whether XAUR is really MAUR 2.0 or a sufficiently different document. Many of the issues are similar to those of media. Treating them in a single document would be preferable, if the issues are sufficiently related.

<Joshue108> JOC: Did you mean XR checklist?

<Joshue108> or XAUR itself?

<Joshue108> JS: I was thinking the XAUR itself.

<Joshue108> <discussion on VR as fake news>

<Joshue108> s/what are the features in XR that make it different from XR?/what are the features in XR that make it different from traditional applications?s\s/what are the features in XR that make it different from XR?\what are the features in XR that make it different from traditional applications?

<Zakim> Joshue, you wanted to ask about follow up with the what is XR definition

<jasonjgw> Josh: we need to schedule a discussion of the defining and unique properties of XR, and their accessiblity implications.

<Joshue108> In WCAG we can't even agree what fully content is!

<Joshue108> s\In WCAG we can't even agree what fully content is!\

Real-time communication accessibility.

jo: Have meeting scheduled with Dom to update him on our work
... We want to know should we go Note? so FPWD? or what would they prefer?

<jasonjgw> Janina: the key concern is related to full support for our real-time text scenarios (including character-by-character transmission and instant messaging-style).

<jasonjgw> Janina notes the regulatory implications.

<jasonjgw> Janina would like this to be clarified at TPAC.

jgw: RTT is definitely an immediately priority, esp in the U.S.
... And, yes, we need to establish good communications to help support a11y in WebRTC going forward

<jasonjgw> Josh will follow up regarding the RTC developmnts.

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.154 (CVS log)
$Date: 2019/08/14 14:00:01 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.154  of Date: 2018/09/25 16:35:56  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/what are the features in XR that make it different from XR?/what are the features in XR that make it different from traditional applications?/
FAILED: s/what are the features in XR that make it different from XR?/what are the features in XR that make it different from traditional applications?s\s/what are the features in XR that make it different from XR?\what are the features in XR that make it different from traditional applications?/
Default Present: jasonjgw, Joshue
Present: jasonjgw Joshue Joshue108
Found Scribe: janina
Inferring ScribeNick: janina
Found Date: 14 Aug 2019
People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)


[End of scribe.perl diagnostic output]